With few exceptions (medical home indicators, difficulty reaching a provider by phone, receipt of all needed care and drugs in Table 6; satisfaction with care in Table 7; and indicators related to homelessness in Table 11), the outcome indicators of interest are available from at least one of the population surveys included in the data scan. A key tradeoff for the analysis of impacts on individuals will be sample size and depth of survey content. The larger surveys that currently allow for state-level estimates--ACS, CPS, and BRFSS--all have limited information on many of the outcome indicators of interest.
For in-depth measurement of topics such as access to services and unmet need that are not addressed in the larger surveys, a few strategies are possible. Multiple years of data could be combined from existing surveys that include these questions (primarily NHIS and MEPS); however, this approach will involve tradeoffs in terms of timeliness of the evaluation and in measuring state-specific impacts. Alternatively, new questions could be added to the existing surveys with large sample sizes, or sample sizes in the existing health surveys could be expanded to allow for more state-specific estimates. MEPS is currently the only survey with measures of patient experience and detailed measures of cost, and it has among the most in-depth measures of health outcomes. Because there are significant time lags in availability of the MEPS data (see Appendix Table A-1a and Table A-1b), the option to combine multiple years of data for the evaluation is one for which tradeoffs in timeliness will need to be carefully considered.
For the extreme poverty group, the tradeoff between survey depth and sample size will be particularly difficult. Even with multiple years of data, the in-depth health surveys will still have fairly limited sample size, and it may be necessary to combine several years of data to perform a meaningful analysis.